(PS4-C56) Examining PTSD Symptom Networks Among Veterans with and Without Positive Traumatic Brain Injury Screen

Introduction: Traumatic brain injury (TBI) has been linked to increased posttraumatic stress disorder (PTSD) prevalence (Hoge et al., 2008) and severity (Ragsdale et al., 2013) among veterans. Despite this, few studies have examined the severity of each individual PTSD symptom among individuals with and without TBI, and none have examined the influence of TBI on the PTSD symptom network; that is, whether the relationships between individual PTSD symptoms differ in those with and without TBI. In the current study, we compared veterans with and without a positive TBI screen on the severity of individual DSM-5 PTSD symptoms. We then computed PTSD symptom networks for these two groups and evaluated differences in the overall connectivity and structure of the networks.

Method: Data was collected as part of a larger study examining the diagnostic accuracy of a PTSD screening tool among consecutive Veterans Health Administration (VA) primary care patients. Participants in the current study (N = 473) were veterans from the larger study with complete data on both the PTSD Checklist for the DSM-5 (PCL-5; Weathers et al., 2013) and a 4-item TBI screen based on the Brief TBI Screen (Schwab et al., 2007). Participants were classified based on the Brief TBI screen into two groups: a TBI group (n = 198) and a non-TBI group (n = 275). Using Bonferroni-corrected t-tests, we compared the TBI and non-TBI groups on mean scores for each of the 20 PCL-5 items. We then computed PTSD symptom networks for the two groups using recommended practices (i.e., Epskamp et al., 2018) and used a permutation hypothesis test called the Network Comparison Test (NCT; van Borkulo et al., 2017) to examine the differences in the structure and connectivity of the two networks.

Results: We hypothesized that individual PCL-5 symptoms would be more severe in the TBI group as compared to the non-TBI group. As expected, each PCL-5 symptom was significantly more severe in the TBI group when compared to the non-TBI group (all ts > 8.11; all ps < .000). However, the NCT showed no significant difference in the structure (M = 0.26, p = 0.22) or in network connectivity (S = 0.39, p = 0.16) of the TBI and non-TBI PTSD symptom networks. Tests of individual edge weights showed that only one edge weight (of 190 possible edges) differed significantly (p = .00) between the two networks: Symptom B1 (intrusive memories) was positively correlated with symptom D2 (negative beliefs) in the TBI group (r = 0.05) but negatively correlated in the non-TBI group (r = -0.05)


Discussion: Findings indicate that although all individual PTSD symptoms were more severe among veterans with a positive TBI screen, the symptoms did not differ in the strength and structure of their connections in a symptom network. We will discuss the implications of these results for future research and will provide additional interpretation of the networks, including discussing the centrality of symptoms within the PTSD networks.